Catabolism of (/?)-Amygdalin and (/?)-Vicianin by Partially Purified ß-Glycosidases from Prunus serotina Ehrh. and Davallia trichomanoides

Gary Kuroki. Pauline A. Lizotte, and Jonathan E. Poulton Department of Botany, University of Iowa, Iowa City, Iowa 52242

Z. Naturforsch. 39 c, 232-239 (1984); received October 3, 1983

Cyanogenic Disaccharides, /?-Glycosidases, Prunus serotina , Davallia trichomanoides , Amygdalin, Vicianin Mature black cherry (Prunus serotina Ehrh.) seeds accumulate high levels of the cyanogenic disaccharide (/?)-amygdalin. Extracts from these seeds contain two /?-glycosidases which have been identified and completely resolved by DEAE-cellulose ion-exchange chromatography. Amygdalin hydrolase hydrolyzed (ß)-amygdalin at an optimum pH of 5.5, releasing (ß)-pruna- sin and D-glucose. This enzyme showed highest activity towards (/?)-amygdalin and failed to hydrolyze (/?)-prunasin. linamarin, /?-gentiobiose and cellobiose. A distinct /?-glycosidase, prunasin hydrolase, displayed a pronounced preference for (/?)-prunasin, hydrolyzing this cyanogenic monosaccharide at an optimum pH of 6.5 to mandelonitrile and D-glucose. Prunasin hydrolase was inactive towards (^)-amygdalin, linamarin, and /?-gentiobiose. Both enzymes showed significant activity towards the artificial substrates /?-ONPGlu and /?-PNPGlu but did not hydrolyze x-PNPGlu. In view of the pronounced specificity of these enzymes towards endogenous cyanogens, it is concluded that upon disruption of black cherry seeds (/?)- amygdalin is catabolized to mandelonitrile in a stepwise manner (the sequential mechanism) by amygdalin hydrolase and prunasin hydrolase with (Ä)-prunasin serving as intermediate. Young fronds of Davallia trichomanoides are rich sources of (Ä)-vicianin (the /?-vicianoside of (/?)-mandelonitrile). A /?-glycosidase, vicianin hydrolase, has been partially purified from frond extracts by ion-exchange chromatography. At the optimum pH of 6.0, this enzyme showed highest hydrolytic activity with (fi)-vicianin, although both (/?)-amygdalin and (/?)-prunasin could be hydrolyzed at approximately 15% of the rate observed with (/?)-vicianin. It failed to hydrolyze /i-gentiobiose, cellobiose, linamarin and x-PNPGlu. Closer examination revealed that (/?)-vicianin and (/?)-amygdalin were hydrolyzed at the aglycone-disaccharide bond (the simultaneous mechanism) yielding mandelonitrile and the respective disaccharides vicianose and /?-gentiobiose.

Introduction aglycone-disaccharide bond would release the

2 -hydroxynitrile and a disaccharide (the “simul­ The catabolism of cyanogenic glycosides is taneous” mechanism). Alternatively, the two sugar initiated by the cleavage of the carbohydrate moiety residues might be removed singly by stepwise by one or more /?-glycosidases [1], Although the hydrolysis with a cyanogenic monosaccharide acting catabolism of several cyanogenic monosaccharides as intermediate (the “sequential” mechanism). In is now well understood [ 2 ], the mode by which the latter case, the two hydrolytic steps might be cyanogenic disaccharides are hydrolyzed remains catalyzed by the same or by distinct /?-glycosidases. largely unclear. Theoretically, two distinct de- Previous reports in the literature [3-6] suggest that gradative pathways may exist for cyanogenic di­ both simultaneous and sequential pathways may saccharides, as shown in Fig. 1. Hydrolysis at the exist in Nature. In order to provide more informa­ tion about this subject, we have selected two cyanogenic plant tissues for closer examination, Abbreviations: PVP, polyvinylpolypyrrolidone; AH, amygdalin hydrolase; PH. prunasin hydrolase; VH, vicia­ namely. Primus serotina seeds and Davallia tricho­ nin hydrolase; /?-PNPGlu and /?-ONPGlu, para- and ortho- manoides fronds. These tissues accumulate high nitrophenyl-/?-D-glucopyranosides; /?-PNPXyl and concentrations of the cyanogenic disaccharides /?-ONPXyl, para- and o/7/;o-nitrophenyl-/?-D-xylopyrano- sides; /ÄPNPGal. p-nitrophenyl-/?-D-galactopyranoside; (/?)-amygdalin (the /?-gentiobioside of (/?)-mandelo- x-PNPGlu. /7-nitrophenyl-x-D-glucopyranoside. nitrile) and (/?)-vicianin (the /?-vicianoside of Reprint requests to Dr. J. E. Poulton. (/?)-mandelonitrile), respectively. Cyanogen-specific 0341-0382/84/0300-0232 S 01.30/0 //-glycosidases have been partially purified from G. Kuroki et al. • Hydrolysis of Cyanogenic Disaccharides 233

P sugar,-0-sugar2 R

N=C—C—0—sugar.-O-sugar,------!---- 2 . ------. NEEC-C-OH Hydroxynitrile ^ ^ + ^ ß-Glycosidase ^ Lyase (_j/

Cyanogenic Monosaccharide Fig. 1. The simultaneous (1) and sequential (II) mechanisms for catabolism of cyanogenic disaccharides.

these plants and their behavior towards endogenous column (1.0x41 cm). The column was eluted with cyanogens and other glycosidic substrates in­ water-saturated «-butanol and fractions containing vestigated. This has allowed us to predict the vicianin were identified by the Feigl-Anger test [ 8 ], pathway by which the endogenous cyanogenic Vicianin was recrystallized from benzene:methanol disaccharides are catabolized upon tissue disrup­ ( 1 : 1, by vol.) and its purity was confirmed by thin tion. layer chromatography. Clorox was purchased from the Clorox Co., Oakland, Ca.

Materials and Methods Plant materials Chemicals Mature black cherry (Prunus serotina Ehrh.) fruits were collected from Hickory Hill Park, Iowa City. Polyvinylpolypyrrolidone (PVP), /?-glucosidase The pits were removed and surface sterilized in 10% (type II, from almonds), simple sugars and chromo- Clorox, blotted dry and stored at 4°C until used. genic substrates were purchased from Sigma Davallia trichomanoides specimens were purchased Chemical Co., St. Louis, MO. Glycosidic substrates from Fountain Square Nurseries, Sacramento, Ca. were available from our laboratory collection. Silica gel K5 TLC plates (thickness, 250 p), 3MM chroma­ Purification and assay of amvgdalin hydrolase and tography paper and DEAE-cellulose were obtained prunasin hydrolase from Prunus serotina from Whatman Chemical Separation Ltd., Kent, U.K., Microcrystalline cellulose (Avicel) was pur­ Unless otherwise indicated, all stages were carried chased from Merck, Darmstadt, BRD. Vicianin was out at 4°C. Black cherry pits (approximately 35) obtained by grinding Vicia angustifolia seeds in were cut in half with a razor blade and the seeds liquid nitrogen with a mortar and pestle. The homogenized with 10 ml of buffer I, 0.5 g PVP, and powdered seeds were placed in boiling 70% 2.0 g glass beads in a mortar. The homogenate was methanol for 10 min and filtered through Whatman filtered through four layers of cheesecloth and grade 1 filter paper. The filtrate was extracted with centrifuged at 17600xg for 25 min. An aliquot an equal volume of petroleum ether. The methanolic (2.5 ml) of the supernatant liquid was chromatog­ fraction was evaporated to dryness, redissolved in raphed on a Sephadex G-25 column (1.5 x 8.3 cm) water, and applied to Whatman 3 MM chromatog­ which had been pre-equilibrated with buffer II. raphy paper. The paper was developed in «-buta­ Elution was carried out with this buffer. The eluate nol: acetic acid:H20 (4:1:5, by vol., upper phase). was applied to a DEAE-cellulose column (1.6x 10cm) After thoroughly drying the paper, the cyanogen which had been pre-equilibrated with buffer II. was located by the Feigl-Anger “sandwich” tech­ After washing the column excessively with buffer II, nique [7] employing the Davallia /?-glycosidase elution of bound proteins was accomplished with a preparation to release HCN. The glycoside was 0 -3 0 0 mM NaCl gradient in buffer II. Fractions eluted with methanol and applied to a cellulose containing amygdalin hydrolase (AH) and prunasin 234 G. Kuroki et al. ■ Hydrolysis of Cyanogenic Disaccharides

hydrolase (PH) activity were pooled separately and strates. the enzyme (0.05 ml) was incubated with

dialyzed against buffer II overnight. These enzyme 1 0 (imol chromogenic substrate (dissolved in 0 . 5 ml preparations were stored at 4 °C and used for assay H 20) and 0.2(imol sodium acetate-HCl buffer, of /?-glucosidase activity as described below. pH 6.0. in a total volume of 1.0 ml. After incubation The standard assay mixture for AH and PH at 30 °C for 1 h. the reaction was terminated by activity contained 3.75 (imol glycosidic substrate adding 2.0 ml of 0.2 M sodium borate-NaOH (dissolved in 0.15 ml H 20). 50 (imol citrate-phos- buffer, pH 9.8, and the developed colors were mea­ phate buffer (pH 5.5 for AH. and pH 6.5 for PH), sured at 400 nm. For other substrates, the standard 30 jal H 20, and 20 (il of diluted enzyme preparation assay mixture for //-glucosidase activity contained (diluted 1:7 with buffer II) in a total volume of 0.5 (.imol glycosidic substrate (dissolved in 20 |al

0.25 ml. After incubation at 30 °C for 5 min, the H 20), 10 (imol sodium acetate-HCl buffer, pH 6.0, reaction was terminated by transferring an aliquot and 20 (il of enzyme in a total volume of 50 jj.1. After (0.23 ml) to a test tube immersed in boiling water. incubation at 30 °C for 1 h. the reaction was After boiling the mixture for 30 s, it was placed on terminated by placing the reaction vial in boiling ice. Aliquots (0.2 ml) were removed and glucose water for 2 min and then on ice. After centrifuga­ production measured by the glucose oxidase tion, aliquots (40 (il) were removed and glucose procedure [9], The assay mixture for chromogenic production determined by the glucose oxidase substrates was essentially the same, but the reaction procedure [9].

was terminated by adding 2.5 ml of 0.2 m sodium Since the hydrolysis of vicianin and amygdalin by borate-NaOH buffer, pH 9.8. The developed colors vicianin hydrolase did not yield glucose (see were measured at 400 nm. Results), the above method was ineffective. Hydrolytic rates for these substrates were therefore Purification and assay of vicianin hydrolase determined in the following manner. The enzyme from Davallia trichomanoides preparation (5-40 (il) was incubated at 30 °C with

All stages were undertaken at 4 °C. Young fronds 1 (imol of glycosidic substrate (dissolved in and fiddle-heads (5 g) of Davallia trichomanoides 40 |il H 20), 10 (imol imidazole-HCl buffer, pH 6.0, were washed with double-distilled water, blotted and 5 (imol of sodium acetate-HCl buffer, pH 6.0, gently dry, and homogenized in 30-40 ml of in a total volume of 0.1 ml. At timed intervals, buffer III, 4 g PVP, and 5.0 g o f glass beads in a aliquots ( 2 0 (il) were removed and subjected to mortar. The homogenate was filtered through four HPLC (Beckman Model 332 Gradient System) on a layers of cheesecloth and centrifuged at 17600 xg reverse-phase Ultrasphere ODS column (4.6 mm x for 20 min. The supernatant liquid was dialyzed 25 cm) using the solvent 30% methanol at a flow against buffer III overnight. The enzyme prepara­ rate of 1.5 ml/min. Hydrolytic rates were calculated tion was then applied to a DEAE-cellulose column from the observed linear decreases in substrate area (1.6x10.5 cm) which had been pre-equilibrated (determined by an Hewlett-Packard 3390A in­ with buffer III. After washing the column with tegrator) with time. buffer III to remove unbound proteins, elution was Reaction products of vicianin and amygdalin accomplished by a linear 0-200 m M NaCl gradient hydrolysis were also co-chromatographed with in buffer III. Fractions containing VH activity, authentic samples of amygdalin, prunasin, /?-gentio- which eluted between 70 and 120 m M NaCl, were biose, D-glucose and L-arabinose on a Whatman pooled and concentrated by ultrafiltration. The TLC silica gel K5 plate. The plate was irrigated enzyme was then applied to a Sephadex G-25 four times using acetonitrile: H20 (85:15, by vol.) column (1.5 x 8.3 cm) which had been pre-equil- as solvent [ 8 ], Compounds were detected by spray­ librated with buffer IV. Elution was undertaken ing the plate with 5% sulfuric acid in methanol with buffer IV, and /?-glucosidase fractions were before heating it at 1 10 °C for 30 min. Composition stored at 4°C until assayed for activity by the of the disaccharides produced by VH activity was following methods. confirmed by subjecting the sugars to hydrolysis in Three assay procedures were employed to deter­ 10% trifluoroacetic acid for lh at 100 °C. The mine the activity of vicianin hydrolase towards the products were analyzed by the above TLC pro­ various potential substrates. For chromogenic sub­ cedure. G. Kuroki et al. • Hydrolysis of Cyanogenic Disaccharides 235

Buffer solutions almonds, almond emulsin and apricots as sources of /y-glycosidases. In sharp contrast to the early conclu­ The following buffer solutions were used: sions of Weidenhagen [18], Haisman and Knight [3] (I) 0 .1 m sodium acetate-HCl buffer, pH 6.0; reported kinetic data suggesting that almond (II) 10 m M histidine-HCl buffer, pH 6.0; (III) 50 m M emulsin contained two (rather than only one) imidazole-HCl buffer, pH 6.0; (IV) 25 m M imi- /?-glucosidases involved in amygdalin hydrolysis. dazole-HCl buffer, pH 6.0. Each enzyme was believed to be specific for one hydrolytic stage. Thus, “amygdalin hydrolase” (AH) Results and Discussion cleaved amygdalin to prunasin and D-glucose, while “prunasin hydrolase” (PH) hydrolyzed prunasin /?-Glycosidases play important roles in the further to mandelonitrile and D-glucose. These degradation [ 2 ] and biosynthesis [ 1 1 ] of many plant enzymes were later identified by continuous electro­ products. Largely through the widespread use of phoresis in a free buffer film [19], but, since their synthetic rather than endogenous substrates during complete resolution and purification to homogeneity their purification and characterization, these were not achieved, their kinetic and molecular enzymes have in past years been regarded as lacking properties rem ain unknown. In 1974, Lalegerie [20] aglycone specificity [12]. This viewpoint has been purified two /?-glucosidase isozymes from sweet recently challenged by the demonstration of several almonds. The non-physiological />-nitrophenyl-/?-D- /y-glycosidases of secondary metabolism having high glucoside (/?-PNPGlu) was used as substrate during specificity for their aglycones [12]. For example, two the purification. These isozymes showed major /i-glucosidases isolated from Triglochin maritima differences when compared with those of Haisman and Alocasia macrorrhiza , plants containing the and Knight. Neither protein displayed absolute cyanogenic monosaccharide triglochinin. exhibited substrate specificity towards either amygdalin or high specificity for this substrate [14, 15]. Other prunasin since each was active towards both sub­ cyanogenic glycosides were hydrolyzed very poorly strates. Furthermore, the observed K m values of or not at all by these enzymes. Likewise, /?-glyco- these cyanogens were 4-5 fold higher than pre­ sidases purified from Sorghum bicolor [16], Linum viously reported [3], In other laboratories, almond usitatissimum [17], and Triglochin maritima [14] emulsin was found to contain three [ 2 1 ] or at least showed great specificity towards the endogenous four [22] /y-glucosidase isozymes. However, in these cyanogens dhurrin, linamarin and taxiphyllin, studies too, the authors used artificial rather than respectively. natural substrates to monitor their enzyme purifica­ Whereas the manner by which cyanogenic mono­ tions. Thus, Helferich and KJeinschmidt [21] and saccharides are hydrolyzed is becoming clearer [ 2 ], Grover et al. [22] in effect achieved the purification no general pattern has yet emerged for the catab­ of /?-PNPGlu- and salicin-hydrolyzing isozymes but olism of the five known cyanogenic disaccharides paid no attention whether these proteins could (i.e. amygdalin, vicianin, lucumin, linustatin and hydrolyze the naturally-occurring cyanogens. The neolinustatin) [3-6]. In view of the pronounced relationship of these proteins to the enzymes in specificity of cyanogenic monosaccharidases, it almonds which hydrolyze amygdalin and prunasin could be reasonably expected that tissues containing still remains unknown. In summary, although it cyanogenic disaccharides should also possess appears likely that amygdalin hydrolysis in almonds /y-glycosidases showing specificity for such com­ occurs in stepwise fashion with prunasin as inter­ pounds. As described in the introduction, these mediate, there is little agreement as to the number compounds could be hydrolyzed by either the and exact nature of /?-glycosidases involved in this simultaneous or sequential routes. These interesting process. questions have been addressed here for the catab­ Employing almond emulsin as the starting-point olism of amygdalin and vicianin. for /y-glycosidase purification may be open to well- founded criticism, since many of the observed The catabolism o f amygdalin in Rosaceous species isozymes could be merely artefacts produced by the The mechanism of amygdalin hydrolysis in modification of parent /?-glycosidases (found in the cyanogenic plants has been investigated using almond seed) during the preparation of almond 236 G. Kuroki et al. • Hydrolysis of Cyanogenic Disaccharides

Fig. 2. DEAE-cellulose chromatography of Prunus serotina homogenate. Chroma­ tography was undertaken as described under Materials and Methods. Fractions were collected and assayed for AH activ­ ity (•—•), PH activity (o— o), and /y-glucosidase activity towards /?-PNPGlu (■----- ■). The protein content was moni­ tored by absorbance at 280 nm (□— □). Fraction emulsin [23, 24], In view of the unavailability of supplied at 15 m M concentration, with the enzymes fresh almonds in Iowa and the uncertainties of at their optimum pH. As shown in Table I, amyg­ utilizing almond emulsin as enzyme source, we have dalin hydrolase showed highest activity towards selected seeds of black cherries (harvested from amygdalin. The synthetic substrates /?-ONPGlu and Prunus serotina trees grown locally) as our source of /?-PNPGlu were hydrolyzed at approximately 50% /?-glycosidases in this study. A crude extract of the rate observed with amygdalin. Slight activity possessing high AH and PH activities was obtained was observed with /?-PNPGal, but the enzyme failed by homogenizing black cherry seeds in sodium acetate buffer as described in the Methods section. Table I. Substrate specificity of Prunus serotina amygdalin Phenolic compounds were removed during homo- hydrolase and prunasin hydrolase. /?-Glycosidase activity genization by PVP and by subsequent gel filtration. was assayed as described in the Methods section using On applying this preparation to a DEAE-cellulose 15 mM concentrations of glycosidic substrates. For com­ parative purposes, hydrolytic rates shown by amygdalin column in 10 m M histidine-HCl buffer, pH 6.0, both hydrolase with these substrates are expressed as a per­ AH and PH activities became bound to the ion- centage of the rate observed with amygdalin as substrate (326 nmol/h/ml enzyme). Hydrolytic rates exhibited by exchanger. Elution of the column with a linear prunasin hydrolase are expressed as a percentage of the 0 -3 0 0 m M NaCl gradient effected the complete rate measured with prunasin (1 540 nmol/h/ml enzyme). resolution of these activities (Fig. 2). Fractions Amygdalin hydrolase Prunasin hydrolase eluting between 65-145 m M NaCl showed high activity towards amygdalin but were inactive Substrate % Activity Substrate % Activity (Amygdalin (Prunasin against prunasin. By contrast, fractions eluting = 100%) = 100%) between 160-235 m M NaCl rapidly hydrolyzed prunasin but were virtually inactive against (^)-Amygdalin 100 (Ä)-Prunasin 100 /?-ONPGlu 57 ß-ONPGlu 79 amygdalin. The activity profile for /?-PNPGlu /?-PNPGlu 46 /?-PNPGlu 18 hydrolysis closely followed those of AH and PH /?-PNPGal 6 Salicin 3 activities throughout the gradient. The AH and PH A-ONPXyl 1 A-PNPGal 1 Maltose 1 Maltose 1 fractions were pooled separately and dialyzed over­ /?-PNPXyl < 1 A-PNPXyl < 1 night to remove sodium chloride prior to the follow­ (/?)-Prunasin 0 Cellobiose < 1 Linamarin < 1 ing investigations. 0 Sucrose /?-Gentiobiose 0 (Ä)-Amygdalin 0 The optimum pH for the hydrolysis of cyano­ x-PNPGlu 0 Linamarin 0 gens was determined using several different buffers. /?-Phenyl- 0 /?-Gentiobiose 0 D-glucoside While AH showed maximum activity at pH 5.5, PH Salicin 0 a-PNPGlu 0 was most active at pH 6.5. More than 50% of the /?-Methyl- 0 /?-Phenyl- 0 maximum rates were realized over the range D-glucoside D-glucoside Cellobiose 0 /?-Methyl- 0 pH 5.0-7.0. The substrate specificities of the two D-glucoside enzyme preparations were investigated by in­ Sucrose 0 /?-PNPXyl 0 cubating various potential glycosidic substrates, Lactose 0 Lactose 0 G. Kuroki et al. • Hydrolysis of Cyanogenic Disaccharides 237

to hydrolyze a wide range of other substrates buffer, pH 6.0, yielded a cell-free extract which including prunasin, linamarin and /?-gentiobiose. released HCN from the endogenous cyanogen Prunasin hydrolase was most active with prunasin vicianin and also from amygdalin and prunasin. but also hydrolyzed /?-ONPGlu and /?-PNPGlu at After dialysis, this preparation was applied to a significant rates. Importantly, this enzyme was DEAE-cellulose ion-exchange column, and unbound inactive towards amygdalin, linamarin, /?-gentio- proteins were removed by washing the column. biose and /?-phenyl-D-glucoside. Bound proteins were eluted by a linear 0-200 m M Thus, our findings with black cherry homogenates NaCl gradient, and fractions were tested for confirm the conclusion of Haisman and Knight [3] /?-glycosidase activity towards vicianin (as judged that AH and PH activities in Prunus species are by the Feigl-Anger test [ 8 ]) and prunasin. As shown catalyzed by two different proteins. It should be in Fig. 3, /?-glycosidase activity towards these sub­ noted that, even at this early stage of purification, strates was eluted under one symmetrical peak by the black cherry /?-glycosidases show surprisingly 6 0 - 130 m M NaCl. The active fractions were pooled narrow substrate specificities. Among the physio­ and investigated more closely after removal of NaCl logical glycosidic substrates tested, each enzyme by gel filtration. shows a pronounced preference for only one cyano­ The substrate specificity of the pooled /?-glycosi- gen and displays little or no activity towards other dase fractions was analyzed by incubating the potential substrates. In view of these strict substrate enzyme preparation with potential substrates (at specificities, we conclude that amygdalin hydrolysis 10 mM concentration) at the optimum pH (pH 6.0). occurs after disruption of black cherries by the As shown in Table II, the HPLC technique in­ sequential mechanism involving stepwise removal dicated that vicianin was overall the best substrate of the glucose residues. Non-physiological chromo- genic substrates were useful in providing additio­

nal information. Both AH and PH hydrolyzed Table II. Substrate specificity of Davallia trichomanoides /7-PNPGlu significantly, preferring this substrate to /?-glycosidase preparation. /?-Glycosidase activity was the corresponding galactose and xylose derivatives. assayed at pH 6.0 as described in the Methods section, using 10 mM concentration of substrates. Reaction rates Their inability to hydrolyze a-glucosides is demon­ were determined in trial 1 by the HPLC technique. In strated by their lack of activity with a-PNPGlu. trial 2, the glucose oxidase assay or a standard spectro- photometric test for the hydrolysis of PNP- and ONP- sugars were employed. Vicianin hydrolases from V. angustifolia and 1 2 Davallia species Substrate Trial Trial % Reaction % Reaction % Vicianin (the /?-vicianoside of (/?)-mandelonitrile) rate Activity rate Activity has so far been detected in only two unrelated (nmol/h/ (Vicianin (|imol/h/ (Prunasin ml enz) = 100%) m lenz) = 100%) genera, namely, Vicia angustifolia (Legum inosae) [4] and several Davallia species (Polypodiaceae) [25], In (/?)-Vicianin 168.4 100 0 0 26.4 16 0.77 3 contrast to the sequential mode of catabolism (/?)-Amygdalin (^)-Prunasin 20.8 12 27.43 100 observed during amygdalin hydrolysis in black /?-ONPGlu - - 6.0 22 cherries, preliminary experiments by Bertrand et al. A-ONPXyl -- 4.41 16 A-PNPGlu —— 2.56 9 [4] and Kasai et al. [5] indicated that vicianin may /?-PNPXyl -- 1.38 5 be catabolized in Vicia angustifolia by the simulta­ /?-PNPGal —— 0.08 0.3 0.05 0.2 neous mode. Crude enzyme extracts from V. an­ Lactose a-PNPGlu <0.01 0 gustifolia seeds liberated the disaccharide vicianose Salicin -— <0.01 0 during vicianin hydrolysis. In our laboratory, we /?-Methyl - -- <0.01 0 have investigated the vicianin-hydrolyzing /?-glyco- D-glucoside Phenyl-/?- -— <0.01 0 sidase from D. trichomanoides to ascertain whether D-glucoside vicianin might be degraded by a similar mechanism Maltose —— <0.01 0 Cellobiose -— <0.01 0 in this taxonomically distinct species. /?-Gentiobiose -- <0.01 0 Homogenization of young Davallia trichomano­ Sucrose -- <0.01 0 <0.01 0 ides fronds with PVP in 50 m M imidazole-HCl Linamarin -- 238 G. Kuroki et al. • Hydrolysis of Cyanogenic Disaccharides 1-10

Fig. 3. Ion-exchange chromatography of Davallia trichomanoides glycosidase on DEAE-cellulose was undertaken as 200 described in the Methods section. r 0.30 Fractions (3 ml) were assayed for /Aglycosidase activity towards vicia- nin (•— •: determined by a semi- 100 2 > quantitative Feigl-Anger test) and 0.15 prunasin (a— a). The protein content was monitored by absorbance at 280 nm (□— □). 10 20 30 40 50 Fraction for the D. trichomanoides enzyme; amygdalin and In conclusion, we believe that upon disruption prunasin were catabolized at only 16% and 1 2 % of of D. trichomanoides fronds vicianin would be the rate observed with vicianin. In a second trial, hydrolyzed by the simultaneous mode, as was hydrolytic rates were determined by the glucose described earlier for the V angustifolia //-glycosi­ oxidase procedure [9] or by a standard spectro- dase. The release of a disaccharide during hydro­ photometric assay for chromogenic substrates [26]. lysis of other secondary plant products has also been

Among those substrates tested in this trial, prunasin reported [ 2 ], was hydrolyzed most rapidly while much slower In summary, the use of endogenous substrates rates were observed with the synthetic glycosides during enzyme purification and characterization has /y-PNPXyl. /y-PNPGlu, /i-ONPXyl and /MDNPGlu. enabled us to demonstrate here the existence of Several other glucosidic substrates including salicin, several cyanogen-specific /?-glycosidases from plant phenyl-/y-D-glucoside, /?-gentiobiose, oc-PNPGlu and the aliphatic cyanogen linamarin were not hy­ drolyzed by this enzyme preparation. The glucose oxidase test revealed that glucose was released only very slowly when the D. trichomanoides /?-glyco- sidase was incubated with amygdalin; this observa­ tion did not agree with the rapid release of HCN detected during this reaction by the Feigl-Anger test. Our suspicion that this enzyme hydrolyses amygdalin by a simultaneous mechanism with release of the disaccharide /i-gentiobiose was con­ firmed by thin-layer chromatography as shown in Fig. 4. Neither prunasin nor glucose, which would be the products expected from a sequential hydrolytic mechanism, were detectable. Likewise, hydrolysis of the endogenous substrate vicianin by the Davallia enzyme yielded only vicianose as the carbohydrate product; moreover, neither prunasin, Fig. 4. Schematic diagram showing the formation of glucose nor arabinose were detectable in the disaccharides upon incubation of the Davallia tricho- manoides ^-glycosidase with amygdalin (Lane 1) and terminated reaction mixture. The identities of vicianin (Lane 2), as detected by the TLC procedure //-gentiobiose and vicianose as reaction products described in the Methods section. The identities of /?-gentiobiose and vicianose were confirmed by subsequent were confirmed by acidic hydrolysis as described in acid hydrolysis and TLC analysis (Lanes 3 and 4. respec­ the Methods section. tively). G. Kuroki et al. ■ Hydrolysis of Cyanogenic Disaccharides 239 tissues containing the cyanogenic disaccharides Acknowledgements amygdalin and vicianin. In Prunus serotina, two This work has been supported in part by Grant distinct proteins, amygdalin hydrolase and prunasin BRSG SO-7 RR7035-17 awarded by the Biomedical hydrolase, are required for the stepwise hydrolysis Research Support Grant Program, Division of of amygdalin by the sequential mechanism. By Research Resources, NIH, to J. E. Poulton and a contrast, the vicianin hydrolase from Davallia tri- Sigma Xi Grant-In-Aid of Research to P. A. Lizotte. chomanoides cleaves vicianin by the simultaneous We also wish to thank the University of Iowa for mode releasing vicianose. The further purification additional funding for HPLC equipment. and characterization of these enzymes are in pro­ gress in this laboratory.

[1] E. E. Conn, in: Ann. Rev. Plant Phys. (W. R. Briggs, [13] W. Hösel and E. E. Conn, TIBS 7 ,2 1 9 -2 2 1 (1982). P. B. Green, and R. L. Jones, eds.), Vol. 31, [14] A W. Nahrstedt, W. Hösel, and A. Walther, Phyto­ pp. 433-451, Academic Press, New York 1980. chemistry 18, 1137- 1141 (1979). [2] W. Hösel, in: The Biochemistry o f Plants: A Com ­ [15] W. Hösel and W. Nahrstedt, Hoppe-Seyler’s Z. prehensive Treatise (E. E. Conn and P. K. Stumpf, Physiol. Chem. 3 5 6 ,1265- 1275 (1975). eds.), Vol. VII, pp. 7 2 5-753, Academic Press, N ew [16] S. H. Eklund, M. S. Thesis, University o f California, York 1981. Davis (1981). [3] D. R. Haisman and D. J. Knight, Biochem. J. 103, [17] G. W. Butler, R. W. Bailey, and L. D. Kennedy, 528-534 (1967). Phytochemistry 4 ,369-381 (1965). [4] G. Bertrand and L. Rivkid, Bull. Soc. Chim. Biol. 4, [18] R. Weidenhagen, Ergebn. Enzymforsch. 1, 197 (1932). 497-499 (1907). [19] D. R. Haisman, D. J. Knight, and M. J. Ellis, Phyto­ [5] T. Kasai, M. Kisimoto, and S. Kawamura. Kagawa chemistry 6, 1501 - 1505 (1967). Daigaku Nogakuba Gakujutso Hokoku 32, 111-120 [20] P. Lalegerie, Biochimie 5 6 ,1297- 1303 (1974). (1981). [21] F. Helferich and T. Kleinschmidt, Hoppe-Seyler’s [6] T. W. M. Fan and E. E. Conn. Plant Physiol. 72, 128 Z. Physiol. Chem. 340, 31-36 (1965). (1983). [22] A. K. Grover, D. MacMurchie, and R. J. Cushley, [7] B. Tantisewie, H. W. L. Rujigrok, and R. Hegnauer, Biochim. Biophys. Acta 482, 9 8 - 108 (1977). Pharm. Weekblad 1 0 4 ,1341 - 1355 (1969). [23] W. D. Loomis, J. D. Lile, R. P. Sandstrom, and [8] F. Feigl and V. Anger, Analyst. 91,282-284 (1966). A J. Burbott, Phytochemistry 18. 1049- 1054 (1979). [9] E. Raabo and T. C. Terkildsen, Scand. J. Clin. Lab. [24] J. W. Anderson, Phytochemistry 7, 1973-1988 (1968). Invest. 12,402-407 (1960). [25] H. Kofod and R. Eyjolfsson, Phytochemistry 8, [10] B. J. White and J. F. Robyt, Proc. Iowa Acad. Sei., 1509-1511(1969). 90, abstract 83 (1983). [26] M. Gross, G. H. Jacobs, and J. E. Poulton, Anal. [11] T. Hemscheidt and M. H. Zenk, FEBS Lett. 110, Biochem. 119,25-30 (1982). 187-191 (1980). [12] S. Veibel, in: The Enzymes (J. B. Sumner and K Myrback, eds.), Vol. 1, pp. 583-620, Academic Press, New York 1950.